summaryrefslogtreecommitdiff
path: root/caffe2/sgd
AgeCommit message (Expand)AuthorFilesLines
2019-03-22Optimize MomentumSGDUpdate maximum block size and make it templatedBilge Acun1-24/+58
2019-01-10Tensor reinitialization codemod - 5/5 (#15884)Jerry Zhang2-10/+10
2018-12-20Tensor construction codemod(ResizeLike) - 7/7 (#15087)Jerry Zhang2-5/+4
2018-12-14FP16MomentumSGDUpdate Op fix and enable for ROCm (#15150)rohithkrn1-3/+3
2018-12-10Tensor construction codemod - 2/3 (#14836)Jerry Zhang2-4/+4
2018-12-10Integrate rocBLAS fp16 api into Caffe2 (#14882)rohithkrn1-1/+1
2018-12-05Shut up "address will always evaluate to 'true'" warnings (#14774)Edward Yang1-1/+1
2018-11-16Change hip filename extension to .hip (#14036)Junjie Bai1-4/+6
2018-11-14device inference for Adam (#13990)Yan Zhu1-0/+9
2018-11-11moving simd adagrad code to perfkernels (#13549)Jongsoo Park1-5/+2
2018-11-07Rename ndim() -> dim() - 6/6Jerry Zhang5-13/+14
2018-11-06Rename ndim() -> dim() - 5/6Jerry Zhang1-1/+1
2018-11-02Rename dim(i) -> size(i) - 2/2Jerry Zhang4-6/+6
2018-10-30Renaming meta() to dtype() - 2/2 (#13334)Jerry Zhang1-1/+1
2018-10-29Renaming size() to numel() - 2/2Jerry Zhang14-115/+115
2018-10-29seperate mkl, mklml, and mkldnn (#12170)Gu, Jinghui1-2/+2
2018-10-24Renaming dims() to sizes() (caffe2/caffe2) - 4/4Jerry Zhang2-2/+2
2018-10-23Guard all Caffe2 protobuf string serializations with CAFFE_ENFORCE (fixed rev...Michael Antonov1-1/+1
2018-10-18Revert D10416438: Guard all Caffe2 protobuf string serializations with CAFFE_...Junjie Bai1-1/+1
2018-10-18Guard all Caffe2 protobuf string serializations with CAFFE_ENFORCE (#12799)Michael Antonov1-1/+1
2018-10-17Serialization takes pointers instead of Blob (#11925)Sebastian Messmer2-3/+5
2018-09-25Fallback CreateMutex/AtomicIter operators for mkl-dnnCheng,Penghui1-0/+9
2018-09-22Remove many caffe2::TIndex and replace them with int64_t (#11943)Christian Puhrsch3-9/+9
2018-09-20codemod: caffe::float16 -> at::Half (#11785)Roy Li4-13/+13
2018-09-20Replace float16 with at::Half in caffe2 (#11676)Roy Li2-15/+12
2018-09-17Support lr adaption for SparseAdam and RowWiseSparseAdam (#11162)Chenguang Xi3-121/+356
2018-09-10Narrowing Blob (#11167)Sebastian Messmer6-11/+11
2018-09-06Windows DLL build with Caffe2 code (#11266)Yangqing Jia1-1/+1
2018-08-31Revert D9554375: Support lr adaption for SparseAdam and RowWiseSparseAdamEdward Yang2-229/+102
2018-08-31Support lr adaption for SparseAdam and RowWiseSparseAdam (#10993)Chenguang Xi2-102/+229
2018-08-23Add CUDA implementation of LARS --caffe2 (#10509)Lin Li3-20/+69
2018-08-16Additional changes to make GPU builds work (#10507)Orion Reblitz-Richardson1-1/+1
2018-08-14Workaround for Cuda9.2 and GCC7 compilation errors (#10510)Mingzhe Li13-67/+67
2018-08-06Add cost inference of fwd sparse operators and sparse adagrad (#9314)Jongsoo Park1-1/+30
2018-08-04Optimize reduce ops for 2d and 3d (#9992)Xiaomeng Yang1-1/+1
2018-08-02Improve lars operator by applying clipping (#9905)Lin Li2-22/+46
2018-07-27Remove template parameter from Tensor (#9939)Jerry Zhang8-35/+40
2018-07-26Revert D8121878: Remove template parameter from TensorJerry Zhang8-40/+35
2018-07-26Remove template parameter from Tensor (#13)Jerry Zhang8-35/+40
2018-07-24Add Adadelta optimizer to caffe2 (#9088)Siddharth Goyal3-0/+424
2018-07-19HIP Operators Generator--> HipOpG (#9322)Peter Yeh7-13/+35
2018-07-13Implementation and operator test for Wngrad optimizer (#8999)Jian Zhang2-0/+311
2018-07-11Fix Eigen issue on OS X with CUDA and nvcc compile (#9350)Orion Reblitz-Richardson2-0/+2
2018-07-10Revert D8768025: [pytorch][PR] Fix Eigen issue on OS X with CUDA and nvcc com...Mike Kelley2-2/+0
2018-07-10Fix Eigen issue on OS X with CUDA and nvcc compile (#9270)Orion Reblitz-Richardson2-0/+2
2018-07-06Add FTRL Optimzier with Group Lasso regularizer (#9074)Xiuyan Ni2-0/+142
2018-06-27fbshipit-source-id: ba600fcd2b5cefc7621357bdeb05e24cea02e5afOrion Reblitz-Richardson1-11/+11
2018-06-26Update from Facebook (#8887)Orion Reblitz-Richardson1-2/+13
2018-06-13Update from facebook (#8384)sf-wind1-1/+1
2018-06-08[cmake] Make cudnn optional (#8265)Yangqing Jia1-3/+0
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-23 01:19:53 +0000